"A problem of rational origin that requires a rational solution." (year 2000 date change problem)(Cover Story)

In the wake of increased press coverage, most information systems professionals and their organizations have become aware of the year 2000 date problem facing the business community. However, most organizations have taken little action on this issue to date.

Not surprisingly, the Gartner Group estimates that one-half of all companies will not be ready when January 1, 2000 arrives. Peter de Jager, an industry consultant, observed in November 1995 that based on his conversations with hundreds of companies, consultants, and vendors, less than 20% of organizations worldwide are addressing the year 2000 date issue.

Even among these, most efforts are at an early stage - focusing on raising problem awareness, assessing the systems affected, and initial planning. Few organizations have embarked on more advanced stages of solving the problem such as revising program code and converting data.

At first glance, particularly to observers outside of the IS community, it appears that the year 2000 date problem was the result of IS departments "dropping the ball." According to this view, IS professionals must have been myopic in utilizing an approach (two-digit year fields) that would eventually lead to systems failure. This paper, however, offers an alternative view of "how we got here," and what is required for an effective solution.

A Problem of Staggering Proportions

Understanding the year 2000 date problem at its most basic level is not difficult. The problem has resulted from the storage of year data in two digits instead of four. This approach has worked for many years since systems could rightfully assume that year data were always preceded by "19." It breaks down, however, at the year 2000 where applications will interpret the year to be 1900 instead of 2000.

This potentially causes problems for any systems that include date references, which is most of them. It is estimated that a typical large organization will pay approximately $40 million to address this problem. Moreover, according to the Gartner Group, fixing the year 2000 date problem will cost about $1 for each line of code in the organization (excluding documentation, training, and final implementation testing), and the total cost worldwide is estimated to be $300 to $600 billion.